The Tokenomics Foundation just published the industry's first large-scale look at how enterprises manage AI economics. Their research covered 472 companies across 11 industries; combined, these companies account for $4.6 trillion in revenue.
One number from the report should stop every AI budget conversation in its tracks:
Within today’s enterprises, teams have gotten reasonably good at seeing what AI costs. They have not gotten good at proving what that spending is worth. Call it the AI value blind spot. The data says this blind spot is not a fringe problem, it's the single largest challenge respondents report having, by a wide margin.
Why are enterprise teams struggling to connect AI token costs to actual business outcomes? This post examines these challenges and shows how teams can begin to eliminate the AI value blind spot.
Confidence in AI spending visibility is decent: Only 20% of respondents say they're not confident in their visibility, and 39% call themselves moderately confident. But ask a harder question, “Can you prove the business outcome that spending produced to your CFO's satisfaction?” and the picture flips. 73% of enterprises are not confident or only slightly confident they can do that. Only 9% are very confident.
What is the difference between AI spending visibility and AI value realization? Visibility and value realization aren't the same skill, and having the former doesn't buy you the latter. Among enterprises that call themselves highly confident in their AI spending visibility, only 25% can still produce a measurable business outcome for their CFO. Among those with merely moderate visibility confidence, 60% can't produce one at all.
| Challenge | Percentage |
|---|---|
| Proving value / ROI | 43% |
| Visibility and attribution of spend | 27% |
| Measurement and data quality | 18% |
| Governance and ownership | 11% |
| Forecasting and unpredictable cost | 11% |
| Skills, literacy and culture | 11% |
| Efficiency and optimization | 9% |
| Model / vendor choice and lock-in | 8% |
| Cost / pricing complexity | 7% |
| Source: Tokenomics Foundation, State of Tokenomics, September 2026. | |
Notice what's at the bottom of that list. Cost and pricing complexity—the thing most AI cost tools are built to solve—is the least commonly cited challenge. Proving value is nearly six times more commonly cited than pricing complexity. Enterprise leaders aren't asking for a cheaper AI bill. They're asking for a defensible answer to what that bill bought them.
That showed up again when the Tokenomics Foundation asked executives what they actually want from their model and token providers. It isn't discounts—that showed up last. Here are the results:
Here's the number that should reset how every enterprise approaches this problem: 88% of enterprises have defined ownership of AI economics somewhere in the organization. Of the 12% that don't, not one of them can connect AI spending to an outcome for the CFO. Zero.
Further, it's not just about having an owner—it's about what that ownership buys you. Enterprises with defined ownership are 3.7x more likely to be able to show the CFO real value from AI investment. Right now, that ownership is scattered: Currently, ownership sits with CTO, CIO, and technology groups in 35% of organizations. In 26% of organizations, ownership is shared across functions. Ownership sits directly with finance in only 5% of organizations. These numbers track, since AI spending has been treated as a technology problem when it's really a business-attribution problem.
As enterprise AI adoption matures, organizations are aggressively turning to optimization tools to bring order to their model infrastructure:"
These numbers clearly demonstrate that enterprises are investing in the infrastructure layer. But routing optimizes what a workload costs to run; it doesn't tell you whether that workload was worth running in the first place. The data backs this up: the enterprises with the best visibility tooling still can't universally answer the value question. Visibility is table stakes. It was never going to be the whole answer.
Everything in this data points to the same missing piece: a direct line from what your organization spends on AI to the work that spending actually produced—the features it shipped, the initiatives it funded, the strategic priorities it was supposed to serve. That line doesn't exist in most enterprises today. That’s because the tools that track AI cost and the tools that track AI's business impact have never lived in the same system.
How can enterprise teams prove the ROI of their AI spending to the CFO? ValueOps AI Tokenomics is built to do just that. It connects AI investment directly to the work behind it—tracing usage to the person or agent running it, to the initiative it was funded under, and to the outcome it produced. That means an AI agent's spending is as visible and as accountable as a person's. Agents get managed alongside your human workforce as a single governed resource, instead of disappearing into an infrastructure bill nobody can trace back to a decision.
That's a capability most vendors in this space haven't built. ValueOps AI Tokenomics is the only solution today built to trace every token consumed—by a person or an agent—back to the strategic initiative and business outcome it was meant to serve. Broadcom is one of the few companies that can provide the full picture required. Our solutions span from the infrastructure layer through to enterprise financial governance. This unique perspective is exactly why we're a founding member of the Tokenomics Foundation, which is helping define this discipline industry-wide.
The real gap is proving the business value AI spending generates, not tracking the costs themselves. According to Tokenomics Foundation data, 73% of enterprises cannot confidently prove AI business outcomes to their CFO despite having solid spending visibility.
AI spending visibility tracks what was spent and where, whereas AI value realization measures whether that spending produced a business outcome worth the cost. The two do not move together—75% of teams with high visibility still fail to prove value to their CFO.
Most enterprises fail to prove AI ROI because spending is not directly attributed to specific strategic initiatives. While visibility tools track costs, 60% of enterprises with clear spending data still cannot measure outcomes because they lack strategic attribution, not dashboards.
No, ValueOps AI Tokenomics complements existing FinOps tools. While FinOps tools track cost and consumption (the "what"), ValueOps AI Tokenomics tracks which strategic investments that spending served and the return it generated (the "why").